Dear LIGGGHTS users,
I tried to insert particles into a cubic simulation box with periodic boundary conditions. The particle sizes vary considerably (Dmax/Dmin=80, D: particle diameter). It is OK to construct the initial packing structure with a core, but when the number of particles become very large and need to use multiple cores, some domains are only filled with large particle with no small particles. I have used "communicate multi" and "neighbor multi" command.
Could you please tell me how to solve this problem?
Any help will really be appreciated.
Best wishes,
June
mschramm | Wed, 12/23/2020 - 05:53
single fill or dynamic fill?
Hello
do you need a dynamic fill while the simulation runs?
For fiber models, I often insert all of my fibers on a single core then run 1 step and write a restart file.
Then I run the rest of my simulation on multiple cores by reading in the previous restart file.
June | Wed, 12/23/2020 - 08:47
I use the command "fix ins
I use the command "fix ins all insert/pack seed ${seed2} distributiontemplate pdd1 vel constant 0 0 0 insert_every once overlapcheck yes all_in yes particles_in_region ${N} region bc"
to insert particles. When I use 4 cores, some domains are only filled with large particles. When using the command "fix ins all insert/rate/region seed ${seed4} distributiontemplate pdd1 vel constant 0 0 -1.0 insert_every 1000 overlapcheck yes nparticles ${N} particlerate 150000 region bc ntry_mc 10000" , large particles are not inserted into the box. It is OK to run the simulation on a single core.
Besides, I want to relax the packing system by the command "minimize 1.0e-4 1.0e-6 100 1000" . The error message：Minimization could not find thermo_pe compute. I tried "compute 1 all pe" and output this compute in thermo output, but still the same error happens.
Do you know how to do energy minimization in LIGGGHTS?
Thank you in advance.
Best wishes,
June